As various mobile devices such as mobile phones are used widely, a usage environment and a usage scenario of a mobile device are further extended. Currently, in many usage environments and usage scenarios, the mobile device needs to collect a voice signal using a microphone of the mobile device.
A mobile device may simply use one microphone of the mobile device to collect a voice signal. However, a disadvantage of this manner lies in that: only single-channel noise reduction processing can be performed, and spatial filtering processing cannot be performed on the collected voice signal. Therefore, a capability of suppressing a noise signal such as an interfering voice included in the voice signal is extremely limited, and there is a problem that a noise reduction capability is insufficient in a case in which a noise signal is relatively large.
To perform noise reduction processing on an audio signal, a technology proposes that two microphones are used to respectively collect a voice signal and a noise signal and perform, based on the collected noise signal, noise reduction processing on the voice signal in order to ensure that a mobile device can obtain relatively high call quality in various usage environments and scenarios, and achieve a voice effect with low distortion and low noise.
Further, to obtain a better spatial sampling feature, a multi-microphone processing technology is further proposed. A principle of the technology is mainly to collect voice signals by separately using multiple microphones of a mobile device, and perform spatial filtering processing on the collected voice signals in order to obtain voice signals with relatively high quality. Because the technology may use a technology such as beamforming to perform spatial filtering processing on the collected voice signals, the technology has a stronger capability of suppressing a noise signal. A basic principle of the technology “beamforming” is that, after at least two received signals (for example, voice signals) are separately processed by an analog to digital converter (ADC), a digital processor uses digital signals output by the ADC to firm, according to a delay relationship or a phase shift relationship between the received signals that is obtained on the basis of a specific beam direction, a beam that points to the specific beam direction.
With improvement in functionality of a mobile device, a current mobile device can work in different application modes, where these application modes mainly include a handheld calling mode, a video calling mode, a hands-free conferencing mode, a recording mode in a non-communication scenario, and the like. Generally, a mobile device that works in different application modes always faces different requirements for a voice signal. However, the foregoing solutions in which a microphone is used to collect a voice signal do not propose how to process the voice signal collected by the microphone to enable a voice signal generated after the processing to meet requirements of the mobile device in different application modes.